IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 27, NO. 6, DECEMBER 1999
1559
Over Recovery in Low Pressure
Sparkgaps—Confirmation of Pressure
Reduction in Over Recovery
K. V. Nagesh, P. H. Ron, G. R. Nagabhushana, and R. S. Nema
Abstract— The over recovery in sparkgaps [1], [2] operating
along the left-hand side of Paschens’ characteristics is due to
pressure reduction in the gap after the first pulse discharge.
This pressure reduction leading to over recovery in low pressure
sparkgaps has been verified using a low pressure sparkgap with
two sparkgaps placed one above the other in the same chamber.
The breakdown voltage strength characteristics of the second
gap has been determined for gap spacings of 3.5 mm to 10
mm, diffusion of plasma in the direction of vacuum pumping
and opposite, at a pressure of 2.1 Pa for hydrogen gas. The
vacuum pumping direction has a great influence on the breakdown strength characteristics of second gap after the first gap
discharge. The breakdown voltage of the second gap can exceed
its self breakdown voltage only 1) when the diffusion of plasma
and vacuum pumping direction are same and 2) when the second
gap spacing is greater than or equal to first gap spacing. Shorter
gaps can always have breakdown voltage lower than or equal to
their self breakdown voltage. The experimental setups, behavior
of self breakdown voltages of second gap due to breakdown in the
first gap, over recovery characteristics of sparkgaps, the results,
and discussions are presented here.
Index Terms— Gas sparkgaps, low pressure discharges, low
pressure sparkgaps, over recovery, pressure reduction in overrecovery, recovery times, sparkgap recovery, sparkgaps.
I. INTRODUCTION
T
HE recovery of the sparkgap between the first and subsequent pulses is a very important phenomenon in repetitive
pulsed sparkgaps. The low pressure sparkgaps recover to
greater than their breakdown voltages of first pulse (FPBDV)
under certain conditions, and it is called over recovery in
these studies [1], [2]. This over recovery is due to pressure
reduction in the gap after the first pulse discharge. The
experiments to measure the pressure in the gap after discharge
are very complicated. However, the changes in pressure after
the discharge can be measured by 1) microwave interferometer
rated 100 GHz [3] or 2) laser interferometer [4] by scanning
the gap volume. Presently, the pressure reduction in low
pressure sparkgaps has been verified using a low pressure
sparkgap with two sparkgaps placed one above the other in
the same chamber and reported here. The breakdown voltage
characteristics of the second gap influenced by the discharge
of the first gap has been determined for gap spacings of 3.5
mm to 10 mm, diffusion of plasma in the direction of vacuum
pumping and opposite, at a pressure of 2.1 Pa for hydrogen gas.
A two pulse system has been used to study these breakdown
voltage characteristics of low pressure sparkgaps with two
gaps. The first pulse is used to initiate discharge in one of the
gaps and a similar second pulse is used to probe the voltage
strength status (after discharge) in the other gap. The first pulse
discharge produces plasma in the sparkgap. This momentarily
increases the gap pressure and temperature to 1.6 10 Pa
[5] and 5000 K [6] in the discharge column. The plasma
expands and diffuses in all direction. The pressure in the gap
volume of the sparkgaps reduces below the normal operating
pressure inversely to the temperature of the plasma. If the
plasma expansion supported by vacuum pumping dominates
the gas inlet, the pressure in the gap volume reduces below
normal pressure and results in over recovery in the same gap.
It results in breakdown voltage of the second gap changes to
greater than its normal self-breakdown voltage. It is assumed
that all ionizations from the discharge in the gap volume are
diffused to the walls by this time. If the temperature and
pressure have been reduced to normal by the time ionizations
are diffused, it results in the full recovery of the same gap.
It results in breakdown voltage of the second gap changes
to nearly the same as its normal self-breakdown voltage.
It results in under recovery for delay times lower than the
above, wherein the pressure in the same gap is higher than
the normal or the leftover ionizations leads to breakdown at
lower voltages. It results in breakdown voltage of the second
gap changes to less than its normal self-breakdown voltage
for periods of up to 10 ms. The effect of gap spacing and
gap pressure on the breakdown voltages of the low pressure
sparkgaps (with single and two gaps) have been studied and
reported here.
II. EXPERIMENTAL SETUP
Manuscript received December 22, 1998; revised September 9, 1999.
K. V. Nagesh and P. H. Ron are with the Accelerator and Pulse Power
Division, Bhabha Atomic Research Centre, Trombay, Mumbai, 400085 India
(e-mail: kvnag@magnum.barc.ernet.in).
G. R. Nagabhushana and R. S. Nema are with the Department of High
Voltage Engineering, Indian Institute of Science, Malleswaram, Bangalore,
560012 India.
Publisher Item Identifier S 0093-3813(99)09853-7.
The cross-sectional view of the low pressure sparkgap with
two gaps is shown in Fig. 1. The experimental setup, the
sparkgap assembly, measurement of voltages and currents are
similar to the one used for over recovery studies with single
sparkgap [1], [2]. The two sets of sparkgaps are assembled
in the same chamber with one set above the other (60 mm
0093–3813/99$10.00  1999 IEEE
1560
IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 27, NO. 6, DECEMBER 1999
Fig. 1. Experimental chamber of low pressure sparkgap with two sparkgaps.
between central axis of electrodes), at the center of polymethyl
methacrylate (PMMA) chamber (150 mm diameter, 100 mm
height). The stainless steel sparkgap electrodes (40 mm diameter, 10 mm thick) are of Rogowskii profile, machined on
a CNC lathe, for a uniform electric field between electrodes.
Two YK-218 cables (65 kV, 20.9 , 250 pF/m, 8 m length
each) are connected to each pair of electrodes
through PMMA insulator and a copper tube for co-axial return
path. These two YK-218 cables are connected to the respective
Tesla transformers at the other ends. All the four cables are
connected to same sparkgap in case of recovery experiments
of a particular sparkgap. The voltage pulses are applied to the
respective sparkgaps through these cables. The outputs of these
sparkgaps are terminated with copper sulphate load of 5.2
High purity Iolar-2 hydrogen gas is used in these experiments.
The gas is let in through the central hole of the electrodes
and
The experimental chamber evacuated to 1.3
10 Pa is filled with the required gas up to 133 Pa pressure
10
Pa before starting the
and evacuated again to 1.3
experiments. This flushing process is repeated four times to
maintain the purity of gas.
(a)
(b)
(c)
III. EFFECT ON BREAKDOWN VOLTAGE
DUE TO FIRST PULSE DISCHARGE
A. Discharge Plasma Diffusion in the Direction of Pumping
Fig. 2 shows the sequence of breakdown and plasma movement processes in a low pressure sparkgap with two sparkgaps,
and
with first pulse applied to the top gap electrode
The typical
second pulse to the bottom gap electrode
Paschen’s characteristic is shown in Fig. 3. The possible ways
is effected are 1) the ultra violet
the breakdowns in gap
illuminates
and
and other lights from the discharge in
hence causes some ionizations which reduces the breakdown
in nanoseconds range, 2) the ions from
voltage of
also enters
and reduces the breakdown voltage of
in
microseconds range, and 3) shockwaves due to discharge in
also causes fluctuation of pressure in the gap
This can
in hundreds
increase or decrease the breakdown voltage of
Fig. 2. Breakdown and plasma movement processes in low pressure sparkgap with two sparkgaps, diffusion in the direction of pumping.
of microseconds to a few milliseconds range. In the present
low pressure sparkgap recovery experiments, the shockwaves
play a major role since the recovery times are in the range
of hundreds of microseconds to a few milliseconds [1], [2].
The shockwave velocity is the combined effect of discharge in
sparkgap and the direction of vacuum system pumping. When
, the velocity is higher
the first pulse is applied to gap
The
compared to the case of first pulse applied to the gap
at a time
and having
first pulse is applied to the gap
This gap breaks down at a voltage
as
a pressure of
shown in Fig. 2(a), (point A on Fig. 3) and produces plasma
NAGESH et al.: OVER RECOVERY IN LOW PRESSURE SPARKGAPS
1561
(a)
(b)
Fig. 3 Typical Paschens’ characteristics of low pressure sparkgap.
(c)
and shockwaves in the gap as shown in Fig. 2(b). This results
in an increase in temperature and pressure in the gaps. The
plasma starts diffusing to the walls resulting in decrease in
temperature and pressure of the gap. The breakdown behavior
of the gaps changes due to diffusion of plasma to the walls.
reduces to
at a time
If the
The pressure in the gap
second pulse is applied at this instant , the gap breaks down
(point on Fig. 3). This shows that the
at a voltage
has been affected due to discharge in
At this time
gap
has not returned to its pre-breakdown status. If
the gap
at a later instant
the second pulse is applied to the gap
with the pressure
(Fig. 2(c), point C on Fig. 3) and
Since there are no
results in breakdown at a voltage
the changes in breakdown voltage
previous discharges in
is possible only with change in pressure in the gap due
to Paschen’s characteristics. This demonstrates that the only
pressure reduction in the gap results in breakdown voltages
higher than the first pulse breakdown voltages after the first
pulse breakdown. When the breakdowns are in the same gap,
the pressure reduction in the gap results in breakdown voltages
higher than the first pulse breakdown voltages during the
recovery period and it is called over recovery in low pressure
sparkgaps.
B. Discharge Plasma Diffusion in the
Opposite Direction of Pumping
Fig. 4 shows sequence of breakdown and plasma movement
in low pressure sparkgap with two sparkgaps, with first pulse
and second pulse to the top
applied to the bottom gap
The first pulse is applied at a time
with a
gap
to the gap
as shown in Fig. 4(a).
pressure in the gap
and produces plasma
This gap breaks down at a voltage
and shockwaves in the gap as shown in Fig. 4(b) (Fig. 3,
point A). This increases the pressure and temperature in the
gaps. This plasma starts diffusing to the walls and the self
breakdown voltage of the second gap starts increasing with
reduction in pressure in the gap. The pressure in the gap
decreases to
at a time
The effect of discharge
Fig. 4. Breakdown and plasma movement processes in low pressure sparkgap with two sparkgaps, diffusion against the direction of pumping.
in
on
are similar to the case mentioned above but
the shockwave velocity can be lower due to shockwave and
vacuum pumping are in the opposite directions. If the second
breaks down
pulse is applied at this instant , the gap
, as shown in Fig. 4(b) (Fig. 3, point
at a voltage
has been affected due to
B). This shows that the gap
and has not returned to its pre-breakdown
discharge in
If the second pulse is applied at a
status up to a time
at which time the pressure is
, as
later instant
shown in Fig. 4(c) (Fig. 3, point D) due to expansion of
the plasma in the gap diffusing against the pumping system,
It can be noted
it may breakdown at a voltage
the
from Paschen’s characteristics that with pressure
However with large
breakdown occurs at a voltage
corresponding to normal
delay times, the pressure reaches
self-breakdown voltage. Breakdowns at higher than normal
self-breakdown voltages are generally not possible in these
cases. Since there are no discharges in the second gap, the
changes in self-breakdown voltages are attributed mainly due
to changes in the gap pressure due to first pulse discharge
corresponding to Paschen’s characteristics.
IV. BREAKDOWN VOLTAGE CHARACTERISTICS
The experiments to determine the breakdown voltage characteristics of low pressure sparkgap with two sparkgaps are
conducted for four gap spacings with hydrogen gas at 2.1
Pa gap pressure. Fig. 5 presents the typical pairs of BDV
and second pulse to gap
pulses of first pulse to gap
Fig. 6 presents the typical waveforms of the current in
The Fig. 7 presents the BDV characteristics of
the gap
after the breakdown of the gap
for different
the gap
1562
IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 27, NO. 6, DECEMBER 1999
Fig. 5. Typical first and second voltage pulse waveforms, (hydrogen gas,
ve polarity, 10 mm gap, 2.1 Pa, 5.2 ):
0
four cables discharge into the sparkgap. This changes the
risetime of the applied pulse as well as current waveshape.
These characteristics for a single low pressure sparkgap have
been presented and discussed in [2].
There are mainly two types of breakdown voltage characteristics in case of two sparkgap setup. In case one, the first pulse
is applied to gap
and in case two, the first pulse is applied to
The BDV’s of gaps
and
have been reduced due
gap
and
, respectively. Initially the
to discharge in the gaps
breaks down after about 2 to 10 s of the
second gap
Once the main discharge
breakdown of the first gap
improves with
is complete, the BDV of second gap
time, increases from 20 to 80% of its predischarge BDV in
200 to 700 s for case one and 27 to 47% for case two. It
reaches prebreakdown conditions in about 1–3 ms time for
with long gap spacings (6.5
case one. The BDV of gap
mm and 10 mm) increases to greater than its predischarge
(3.5
BDV in less than 10 ms. But the BDV of short gap
mm) does not reach its predischarge BDV even up to 10 ms.
It may be noted that each gap recovery characteristics shows
over recovery when tested independently (Figs. 9 and 10).
V. DISCUSSIONS
Fig. 6. Typical current pulse waveforms, hydrogen gas,
mm gap, 2.1 Pa, 5.2 , 50% current reversal.
0ve polarity, 10
gap spacings. The Fig. 8 presents the BDV characteristics of
after the breakdown of the gap
and vicethe gap
versa for 10 mm gap spacings. Each of the characteristics of
Figs. 7 and 8 are the average of four similar characteristics.
have been recorded at
The self BDV’s of the gaps
regular intervals during the BDV characteristics experiments
which are plotted in these characteristics. The recordings show
breakdown voltages of the second gap higher than the first gap
of the low pressure sparkgap with two sparkgaps. Figs. 9 and
10 shows the recovery characteristics of the two sparkgaps
along with spread in breakdown voltage data, after the first
pulse breakdown. These characteristics show over recovery
for both sparkgaps in the range of 25–30%. It may also
be noted that the breakdown voltages in these experiments
are low compared to two gap experiments due to higher
load capacitance (4000 pF against 2000 pF for two gaps).
In case of two sparkgaps, two cables discharge through the
sparkgap. In case of recovery experiments of single gap, the
In case of low pressure sparkgap with two sparkgaps, there
is a small difference in self BDV of these two gaps due to
statistical variations. However, it is not significant. The BDV
reduces approximately to zero voltage in less
of gaps
than 5 s–10 s (Figs. 7 and 8) due to fast diffusion of plasma
,
at a velocity of approximately 1–2 cms/ s from gaps
improves slowly
respectively. The BDV of the gaps
with time due to diffusion of ions to the walls. It increases to
about 1 kV in 25 s in case one (Fig. 7) and about 2 kV in
25 to 60 s in case two (Fig. 8). This shows that majority of
in 25 s with an
the ionizations have moved to the gap
cms/ s for case one
average velocity of 6 cms/25 s
in 60 s with an average velocity of 6
and to the gap
cms/ s for case two due to shockwave by
cms/60 s
this time. Once the ionization from first gap diffusing to the
second gap decrease to zero, the second gap returns to normal
after the discharge in gap
at a faster rate. The BDV’s of
are higher than self BDV of the gap
for gap spacings
of 10 mm, 6.5 mm, and 5 mm, and lower for 3.5 mm for case
does not increase to greater than
one. The BDV of gap
or equal to its predischarge BDV for case two. This indicates
does not reduce to below its
that the pressure in the gap
normal pressure due to slow moving shockwave. Since there
change in BDV
is no previous discharge in the gaps
to greater than its self-breakdown value is possible only with
reduction of pressure in the first gap due to discharge in the
second gap. This indicates that the pressure in the gap
reduces below normal pressure due to the discharge in the
for 5 mm, 6.5 mm, and 10 mm gap spacings. The
gap
may completely enter the larger gap
shockwave of the gap
and reduce the pressure in the gap to below
spacings of
normal gap pressure. The shock wave may not enter to the
without being deflected by the electrodes in case of
gap
3.5 mm gap spacings. The increase in BDV’s of 6.5 mm and
NAGESH et al.: OVER RECOVERY IN LOW PRESSURE SPARKGAPS
1563
0
Fig. 7. Second pulse BDV characteristics, hydrogen gas, ve polarity, 50% reversal (1) SBV of G2 = 42 kV, (2) SBV of
G2 = 39 kV, (4) G1 = 5 mm, G2 = 6:5 mm, (5) G1 = 5 mm, G2 = 5 mm, and (6) G1 = 5 mm and G2 = 3:5 mm.
Fig. 8. Second pulse BDV characteristics, hydrogen gas,
pulse to G1, and (4) first pulse to G2:
0ve polarity, 50% reversal (1) SBV of
10 mm gap spacings to greater than their self BDV’s clearly
demonstrates that the pressure in the gap reduces to lower
G
G
2 = 40 kV, (2) SBV of
2 = 40 kV, (3) SBV of
G
1 = 39 kV, (3) first
than its normal pressure, since there is no previous discharge
in this gap. This also indicates that the probability of change in
1564
IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 27, NO. 6, DECEMBER 1999
Fig. 9. Significance of over recovery with range of breakdown voltages of first gap for hydrogen gas, 5 mm gap,
5.2 load, 2.1 Pa pressure: (1) recovery pulse BDV’s and (2) first pulse BDV’s.
Fig. 10. Significance of over recovery with range of breakdown voltages of second gap for hydrogen gas, 5 mm gap,
5.2 load, 2.1 Pa gap pressure: (1) recovery pulse BDV’s and (2) first pulse BDV’s.
BDV to greater than its self-BDV value in one sparkgap due
to the discharge in the other is very high when the shockwave
has to move in the direction of pumping system. The pressure
0ve polarity, 50% current reversal,
0ve polarity, 50% current reversal,
reduction in the gap below normal pressure after discharge
is the main reason for over recovery of the low pressure
sparkgaps to the left of Paschen’s minimum.
NAGESH et al.: OVER RECOVERY IN LOW PRESSURE SPARKGAPS
VI. CONCLUSIONS
The discharge in one of the gaps reduces the BDV of the
other gap placed some distant apart in a two gap setup. This
change is a function of shockwave velocity and direction of
pumping. This can lead to changes in BDV’s, greater than its
predischarge BDV of the other gap. The change in BDV’s is
less predominant in gaps greater than or equal to the first gap.
The over recovery of the low pressure sparkgaps can result
when the shock wave produced due to first pulse discharge
can reduce the pressure in the same gap to below normal
pressures. Even though the experiments have been conducted
for only hydrogen gas, similar behavior can be expected from
other gases like argon, deuterium, etc.
ACKNOWLEDGMENT
The authors are thankful to staff members of Accelerator and
Pulse Power Division for useful discussions and support. They
would also like to thank Dr. S. S Kapoor, Director, Physics and
E & I Groups, BARC, for his keen interest and encouragement.
REFERENCES
[1] K. V. Nagesh, “A study of recovery times of low pressure sparkgaps,”
Ph.D. dissertation, Indian Inst. Sci., Bangalore, India, July 1997.
[2] K. V. Nagesh, P. H. Ron, G. R. Nagabhushana, & R. S. Nema, “Over
recovery in low pressure sparkgaps,” IEEE Trans. Plasma Sci., vol. 27,
pp. 199–210, Feb. 1999.
[3] H. Hermansdorfer, “Microwave diagnostics techniques,” in Plasma Diagnostics, W. Lochte-Holtgreven, Ed. Amsterdam, The Netherlands:
North Holland, 1968, p. 478, ch. 8.
1565
[4] S. M. Turnbull, S. T. Macgregor, F. A. Tuema, and O. Farish, “A
quantitative laser Schliren method for measurement of neutral gas
density in high pressure gas switches,” Meas. Sci. Technol., vol. 4, pp.
1154–1159, 1993.
[5] G. A. Mesyats and D. I. Proskurovsky, Pulsed Electrical Discharges in
Vacuum. New York: Springer-Verlag, July 1988, ch. 4.53,
[6] K. Tsuruta and H. Ebara, “Modeling of a gas temperature decay after
arc discharge in small air gaps,” IEEE Trans. Dielect. Elect. Insul., vol.
27, pp. 451–456, June 1992.
K. V. Nagesh, for a photograph and biography, see p. 210 of the February
1999 issue of this TRANSACTIONS.
P. H. Ron received the B.E. degree from Poona
University, India, in 1961, the M.Sc.(Engg.) degree
from the University of Manchester Institute of Technology, Manchester, U.K., in 1969, and Ph.D degree
from the Indian Institute of Science, Bangalore,
India, in 1984.
He was a Visiting Scientist at the National Research Council, Ottawa, Ont., Canada, in 1985 and
McGill University, Montreal, P.Q., Canada, in 1987.
He has been working in the fields of high voltage
insulation engineering, multigigawatt pulse power
systems, pulsed high magnetic fields, EMI/EMC, and industrial electron
accelerators.
R. S. Nema, for a photograph and biography, see p. 210 of the February
1999 issue of this TRANSACTIONS.